15:00:11 #startmeeting neutron_l3 15:00:12 Meeting started Thu Dec 17 15:00:11 2015 UTC and is due to finish in 60 minutes. The chair is carl_baldwin. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:00:14 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 15:00:17 The meeting name has been set to 'neutron_l3' 15:00:22 hi 15:00:23 #topic Announcements 15:00:26 o/ 15:01:11 The release schedule page has moved 15:01:13 #link http://docs.openstack.org/releases/schedules/mitaka.html 15:01:43 We're somewhere around the middle of Mitaka-2 15:02:20 Considering that many will be taking vacation around Christmas and New Years, it is coming up quickly. 15:02:41 Anything else to announce? 15:03:25 #topic Bugs 15:03:32 mlavalle: ping 15:03:34 hi 15:03:53 we have the same 2 high importance bugs as last week plus one 15:04:06 first one is https://bugs.launchpad.net/neutron/+bug/1478100 15:04:06 Launchpad bug 1478100 in neutron "DHCP agent scheduler can schedule dnsmasq to an agent without reachability to the network its supposed to serve" [High,In progress] - Assigned to Cedric Brandily (cbrandily) 15:04:48 Cedirc pushed next PS to proposed solution and is awaiting reviews: https://review.openstack.org/#/c/205631/ 15:05:33 any other comments on this one? 15:05:50 I'll review that one if I can figure out how to read the new gerrit patch review page. 15:06:05 * carl_baldwin thinks someone hit it with an ugly stick 15:06:12 carl_baldwin: yeah, I got confused last night as well :-) 15:06:32 next up is https://bugs.launchpad.net/neutron/+bug/1516260 15:06:32 Launchpad bug 1516260 in neutron "L3 agent sync_routers timeouts may cause cluster to fall down" [High,In progress] - Assigned to Oleg Bondarev (obondarev) 15:07:18 amuller has a fix proposed here https://review.openstack.org/#/c/245432/ that has been -1 recently so it needs next revision 15:07:55 and obondarev has another solution proposed https://review.openstack.org/#/c/234067/. This one is awaiting review 15:08:40 I would like clarification about the relationship between the two proposed fixes. I'll add a comment to the bug report. 15:09:22 there's been some discussion between amuller and obondarev in the Gerrit. Please see the comments.... 15:09:48 Finally, we have a new one https://bugs.launchpad.net/neutron/+bug/1524916 15:09:48 Launchpad bug 1524916 in neutron "neutron-ns-metadata-proxy uses ~25MB/router in production" [High,New] 15:10:22 I went through the comments last night and it wasn't clear to me whether we really think this is an issue or not 15:10:53 I pinged ajo earlier today, but hasn't responded. I'll follow up with him 15:11:20 any comments or insights? 15:12:25 if not, I want to mention that I am going through the list of bugs, purging and updating 15:12:51 there is a number of old reports without activity for a long time that I am marking incomplete 15:12:56 haleyb: Do you recall looking at memory usage of the metadata proxy in HP public cloud? I have a vague recollection. 15:13:48 carl_baldwin: i think so, since we're spinning-up so many. i think kuba (red hat) or ajo did too 15:14:35 i didn't see any easy solution since part of it was the modules it imported 15:15:05 It never seemed to be a High importance problem for us. We had other problems with 100s of routers on the network node before that. 15:15:25 I'm considering bumping it down in priority. Maybe I'll suggest that on the bug report and see what the reaction is. 15:16:33 carl_baldwin: that is all I have as far as bugs 15:16:38 mlavalle: Thanks for your work on scrubbing the old bugs. 15:17:05 * carl_baldwin is the Neutron bug deputy this week. Thinking of trying to do the same with the overall Neutron bugs. 15:17:54 mlavalle: Great work! 15:18:03 :-) 15:18:11 #topic Routed Network Segments 15:18:42 The demand for this continues to grow. I think this is going to be a very important project for Neutron. 15:18:54 However, consensus still seems distant. 15:19:33 I'm currently discussing with armax and kevinbenton (approver for this blueprint) to explore options and reach consensus before releasing the next version of the spec in to the wild. 15:20:05 If that goes well, it could be a huge step toward getting consensus. 15:20:42 Anyway, we're going to need developers to work on this once it is achieved. 15:20:54 That's all. 15:21:05 * mlavalle want to pitch in 15:21:37 wants^^^ 15:21:55 carl_baldwin: I m in 15:21:58 mlavalle: It will be great to have your help. 15:22:04 vikram_: :) 15:23:22 #topic bgp-dynamic-routing 15:23:30 vikram_: ping 15:23:42 I think tidwellr is out for the rest of the year on vacation. 15:23:44 carl_baldwin: pong 15:24:02 he must be enjoying the baby 15:24:04 carl_baldwin: I am working for the DVR stuff now 15:24:52 carl_baldwin: Need some review help for Agent, Scheduler and Driver patch 15:25:21 carl_baldwin: These patches are almost done.. One round of review would be helpful 15:25:56 vikram_: i can review your patch 15:25:59 vikram_: From the neutron meeting on Tuesday, it sounds like the merges conflicts are scaring people away. 15:26:12 carl_baldwin: Will resolve it 15:26:19 vikram_: Could you link the specific patches you want reviewed here? 15:26:20 carl_baldwin: and make it clean .. 15:26:40 carl_baldwin: 1 sec 15:27:49 https://review.openstack.org/#/c/207607/ 15:28:14 https://review.openstack.org/#/c/207625/ 15:28:20 https://review.openstack.org/#/c/207635/ 15:28:54 that's all 15:29:16 juno_zhu: thanks;) 15:29:38 vikram_: Thank you. I will answer your call to review this week. 15:29:47 carl_baldwin: Thanks ;) 15:29:55 * carl_baldwin really hopes he can figure out that new review page on gerrit. ;) 15:30:27 vikram_: why not create a new review patch including all the code changes and then review the new patch 15:30:49 juno_zhu: change is huge 15:31:14 vikram_: you mentioned in last meeting that bgp dynamic routing would be moved from service plugin to ml2 15:31:18 juno_zhu: so the code is broken in pieces for effective review 15:31:35 vikram_: you have not updated the code about this change 15:31:38 juno_zhu: it already moved 15:31:47 juno_zhu: it's ml2 now 15:32:02 vikram_: which patch? 15:32:12 juno_zhu: the patches I requested for review is related to agent, scheduler and driver 15:32:30 juno_zhu: CRUD patch needs Ryan attention ;) 15:33:12 juno_zhu: provided the link above.. 15:33:27 vikram_: thanks 15:33:49 juno_zhu: https://review.openstack.org/#/c/201621/ 15:33:55 juno_zhu: CRUD patch 15:33:55 #topic dns-resolution 15:34:02 hi 15:34:51 I got https://review.openstack.org/#/c/212213/ reviewed late last week. armax proposed improving it decoupling functionality from ml2 15:34:55 mlavalle: This one was fun this week. :) 15:35:06 yeah it is 15:35:36 so I am re-coding along the ml2 extension manager line that he suggested 15:36:11 I am almost done. I should be done today and will start testing it again today. I expect to have it ready for review again at trhe end of this week 15:36:40 the imporvements are what carl_baldwin and I discussed on Monday 15:36:51 mlavalle: Great, ping me when ready and I'll review. 15:37:16 that's all I have this week. Thanks to carl_baldwin and armax for the help 15:37:35 mlavalle: Still, good progress here. It is cool to see the functionality coming together. 15:37:54 #topic Address Scopes 15:38:10 yeah it's completely funtional and bug free. The discussion now is on implementation improvement 15:38:42 This topic has seen some good progress this week. 15:39:22 The next review that needs to merge is https://review.openstack.org/#/c/189500/ 15:39:46 I need to decide if the RPC version should really be bumped for this. 15:40:30 Part of me wants to not bump it because either server or agent side will still work if the other side gets upgraded. It almost feels like bumping the version number of RPC will cause more disruption. 15:40:44 haleyb: Any thoughts here? 15:40:56 * haleyb wakes up 15:42:25 i am thinking gary is right 15:42:53 Other than that, there is a neutron client patch that should be merging now (has +A) but is not. I'll check in to that. 15:43:37 And, I need to get the L3 agent patches in to a devstack to pull them back together. 15:43:50 Thanks to xiaohhui for a lot of help here. 15:44:25 np:) carl 15:46:02 haleyb: Could you note your position on the review? 15:46:29 That is all I have for address scopes. I hope to get these patches in to devstack early next week and get them in to shape. 15:46:33 will do 15:46:52 #topic Open Discussion 15:47:02 Any other topics that we should tackle? 15:47:33 does any one know if the registration for talks at Austin summit has opened? 15:48:08 i hope not :) 15:48:46 I'm pretty sure it hasn't opened. 15:50:46 Well, here's 10 minutes back. Don't spend it all in one place. :) 15:50:49 #endmeeting